Docstoc

Introduction

Document Sample
Introduction Powered By Docstoc
					                    Introduction

• MPEG, Moving Picture Experts Group was started in
  1988 as a working group within ISO/IEC with the aim
  of defining standards for digital compression of audio-
  visual signals.
• MPEG’s first project, MPEG-1 was published in 1993
  as ISO/IEC 11172.
• In general MPEG is asymmetrical process: it has
  been estimated that encoding needs 100 times more
  processing power compared to the decoding process.


Introduction               tMyn                         1
• MPEG-1 consists of six parts:

    system                ISO/IEC 11172-1
    video                 ISO/IEC 11172-2
    audio                 ISO/IEC 11172-3
    low bit rate audio    ISO/IEC 11172-4
    conformance testing   ISO/IEC 11172-5
    simulation software   ISO/IEC 11172-6




Introduction              tMyn              2
• ”Coding of moving pictures and associated audio for
  digital storage media”
• The original goal was playback of digital audio and
  video using a standard CD with a bit rate of 1.416
  Mbits/sec.
• MPEG 1 was designed to handle progressive (also
  referred to as noninterlaced) video.
• MPEG1 uses a family of three audio coding
  schemes, called Layer 1, Layer 2 and Layer 3, with
  increasing complexity and sound quality.
• The three layers are hierarchical: a Layer 3 decoder
  handles Layers 1, 2 and 3 and so on…

Introduction              tMyn                           3
• All layers support 16-bit digitized audio using 16 (only
  MPEG 2), 22.05 (only MPEG 2), 24 (only MPEG 2),
  32, 44.1 or 48 kHz sampling rates.
• Layers work typically with bit rates:
     – Layer1: 32-448 kbits/sec
     – Layer 2: 8-384 kbits/sec
     – Layer 3: 8-320 kbits/sec

• Two audio channels are supported with four modes
  of operation:
     –   normal stereo
     –   joint(intensity and/or ms) stereo
     –   dual channel mono
     –   single channel mono


Introduction                       tMyn                  4
• For normal stereo, one channel carries the left audio
  signal and one channel carries the right audio
  channel.
• For intensity stereo (supported by all layers) high
  frequencies (above 2 kHz) are combined. The stereo
  image is preserved but only the temporal envelope is
  transmitted.
• For ms stereo (supported by Layer 3), one channel
  carries the sum signal (L+R) and the other the
  difference (L-R) signal.




Introduction              tMyn                            5
• To determine which layer should be used for a
  specific application, look at the available bit rate, as
  each layer was designed to support certain bit rates
  with a minimum degradation of sound quality.
• Layer 1, a simplified version of Layer 2, has a target
  bit rate of about 192 kbits/sec per channel or higher.
• Layer 2 is identical to MUSICAM, and has a target bit
  rate of about 128 kbits/sec per channel.
     – Masking-pattern adapted Universal Subband Integrated
       Coding and Multiplexing




Introduction                   tMyn                           6
• Layer 3 (also known as mp3) merges the best ideas
  of MUSICAM and ASPEC and has a target bit rate of
  about 64 kbits/sec per channel.
     – Adaptive Spectral Perceptual Entropy Coding




Introduction                   tMyn                   7
• All layers use a coding scheme based on
  psychoacoustic principles – in particular, ”masking”
  effect where, for example, a loud tone at one
  frequency prevents another, quieter, tone at a nearby
  frequency from being heard.
• Adjusting the noise floor reduces the amount of
  needed data, enabling further compression.
• CDs use 16 bits of resolution to achieve a signal-to-
  noise ratio (SNR) of about 96 dB, which just happens
  to mach the dynamic range of hearing pretty well.




Introduction              tMyn                        8
• If 8-bit resolution were used, there would be a
  noticeable noise during silent moments in the music.
• However, noise isn’t noticed during loud passages
  due to the masking effect, which means that around a
  strong sound you can raise the noise floor since the
  noise will be masked anyway.
• For a stereo signal, there usually is redundancy
  between channels.




Introduction             tMyn                        9
• Table 1 lists some of the constrained parameters.
• The CPB (Constrained Parameters Bitstream) is a
  limited set of sampling and bit rate parameters
  designed to standardize buffer sizes and memory
  bandwidths, allowing a nomimal guarantee of
  interoperability for decoders and encoders, while still
  addressing the widest possible range of applications.




Introduction               tMyn                         10
           horizontal resolution     max 768 samples (pixels)

           vertical resolution       max 576 scan lines

           picture area,
                                     max 396 macroblocks
           number of macroblocks
           pel rate,
                                     max 396*25 macroblocks per second
           number of macroblocks/s

           picture rate              max 30 frames per second

           bit rate                  max 1,856 Mbps




           Table 1. Some of the constrained parameters for MPEG 1.




Introduction                                 tMyn                        11
• The CPB limits video to 396 macroblocks (396*16*16
  = 101376 pixels).
• Therefore, MPEG 1 video is typically coded at SIF
  resolutions of 352*240 or 352*288.

• SIF, Standard Input Format:
  352 (horizontal)*240 (vertical)*29,97 (fps), NTSC
  352 (horizontal)*288 (vertical)*25,00 (fps), PAL




Introduction              tMyn                        12
• MPEG 1 video data uses the 4:2:0 YCbCr format
  shown in Figure 1.




Introduction            tMyn                      13
     Active line
     number

            1



            2



            3



            4



            5



            6



                      = Y sample

                       = Calculated Cb, Cr sample


            Figure 1. Sampling for 4:2:0 YCbCrMPEG 1.

Introduction                                   tMyn     14
• MPEG 2, ”Generic coding of moving pictures and
  associated audio information” , extends MPEG 1 to
  cover a wider range of applications.
• The primary application targeted during the definition
  process was all-digital transmission of broadcast-
  quality video at bit rates of 4-9 Mbits/sec.
• MPEG 2 is an ISO/IEC 13818 standard, and consists
  of nine parts:




Introduction              tMyn                         15
    systems               ISO/IEC 13818-1
    video                 ISO/IEC 13818-2
    audio                 ISO/IEC 13818-3
    conformance testing    ISO/IEC 13818-4
    software simulation    ISO/IEC 13818-5
    DSM-CC extensions      ISO/IEC 13818-6 (Digital
    Storage Media Command and Control)
    advanced audio coding ISO/IEC 13818-7 (Dolby
    Digital 5.1)
    RTI extension          ISO/IEC 13818-9 (Real
    Time Interface)
    DSM-CC conformance ISO/IEC 13818-10

Introduction              tMyn                        16
• Figure 1 illustrates the positioning of YCbCr samples
  for the 4:2:0 format.
• 4:2:0 YCbCr implements a 2:1 reduction of Cb and Cr
  in both the vertical and horizontal directions.
• To display 4:2:0 YCbCrdata, it is first converted to
  4:4:4 YCbCr data, using interpolation to generate the
  new Cb and Cr samples.
• Actually MPEG 2 4:2:0 is not same as MPEG 1, see
  the spatial relationships between luminance and color
  difference samples.




Introduction              tMyn                       17
     Active line
     number

            1



            2



            3



            4



            5



            6



                       = Y sample

                       = Calculated Cb, Cr sample


            Figure 1. 4:2:0 YCbCr sampling forMPEG 2. The sampling positions on the active scan
            lines of a progressive or noninterlaced picture.

Introduction                                   tMyn                                         18
• Figure 2 illustrates the positioning of YCbCr samples
  for 4:2:2 format.
• For every two horizontal Y samples, there is one Cb
  and one Cr sample.
• Each sample is typically 8 bits (consumer
  applications) or 10 bits (pro-video applications) per
  component.
• Each sample therefore requires 16 bits (or 20 bits for
  pro-video applications), usually formatted as shown
  in Figure 3.




Introduction               tMyn                        19
     Active line
     number

            1



            2



            3



            4



            5



            6



                       = Y sample

                       = Calculated Cb, Cr sample


            Figure 2. 4:2:2 YCbCrsampling for MPEG 2. The sampling positions on the active scan
            lines of a progressive or noninterlaced picture. Co-sited sampling.

Introduction                                   tMyn                                         20
        SAMPLE         SAMPLE         SAMPLE         SAMPLE
        0              1              2              3

          Y7-0           Y7-1           Y7-2           Y7-3
          Y6-0           Y6-1           Y6-2           Y6-3
          Y5-0           Y5-1           Y5-2           Y5-3
          Y4-0           Y4-1           Y4-2           Y4-3
          Y3-0           Y3-1           Y3-2           Y3-3
          Y2-0           Y2-1           Y2-2           Y2-3
          Y1-0           Y1-1                                  -0 = SAMPLE 0 DATA
                                        Y1-2           Y1-3
          Y0-0           Y0-1                                  -1 = SAMPLE 1 DATA
                                        Y0-2           Y0-3
                                                               -2 = SAMPLE 2 DATA
                                                               -3 = SAMPLE 3 DATA
          CB7-0         CR7-0          CB7-2           CR7-2
          CB6-0         CR6-0          CB6-2           CR6-2
          CB5-0         CR5-0          CB5-2           CR5-2
          CB4-0         CR4-0          CB4-2           CR4-2
          CB3-0         CR3-0          CB3-2           CR3-2
          CB2-0         CR2-0          CB2-2           CR2-2
          CB1-0         CR1-0          CB1-2           CR1-2
          CB0-0         CR0-0          CB0-2           CR0-2



      Figure 3. Frame buffer formatting in MPEG 2 4:2:2.

Introduction                                 tMyn                          21
• For broadcast applications sufficient chrominance
  resolution can be provided with the 4:2:0 sampling.
• This method of sampling is generally accepted as the
  most applicable level and profile for MPEG 2, and is
  termed Main Profile at Main Level or MP@ML.
• Figure 4 illustrates MPEG 2 4:2:0 sampling in Digital
  TV environment.




Introduction              tMyn                       22
Active line
number               FIELD N                                      FIELD N+1


      1

      [2]

       3

      [4]

       5

      [6]

       7

      [8]




                          = Y sample                     [] = Field N+1

                          = Calculated Cb, Cr sample


               Figure 4. Sampling 4:2:0 YCbCrfor MPEG 2. The sampling positions on
               the active scan lines of an interlaced picture (top_field_first = 1).

    Introduction                                  tMyn                                 23

				
DOCUMENT INFO
Shared By:
Categories:
Stats:
views:4
posted:9/10/2011
language:English
pages:23